Metadata-Version: 2.4
Name: lmrs-client
Version: 0.1.29
Summary: Client library for the Local Model Runtime Service (LMRS)
Author-email: Vasiliy Zdanovskiy <vasilyvz@gmail.com>
License: MIT
Requires-Python: >=3.12
Description-Content-Type: text/markdown
Requires-Dist: mcp-proxy-adapter>=8.11.1

# lmrs-client

Client library and CLI for the Local Model Runtime Service (LMRS).

Each client method mirrors one public LMRS command. Transport, TLS and error
mapping are handled by the client; callers never build URLs or JSON-RPC payloads.

## Install

    pip install lmrs-client

## Usage

    import asyncio
    from lmrs_client import LmrsClient

    async def main():
        client = LmrsClient(host="lmrs", port=8012, protocol="https", ca="/etc/lmrs/ca.crt")
        info = await client.current_model()
        print(info["model"], info["context_window"])

    asyncio.run(main())

`current_model()` reports the model the running vLLM serves and its context window.

## Command line

    lmrs-client --help
    pipeline --help

`pipeline` is the fleet-contract acceptance runner for a deployed LMRS server.
